Search: “reasoning”

SLIP: Segmentation with Low-latency Interactive Prompting for 3D Medical Images

arXiv:2607.22332v1 Announce Type: new Abstract: Interactive deep image segmentation enables efficient medical image annotation by iteratively refining predictions from user prompts, such as positive a

researcharxiv-cs-cv

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Model Releases

Small context windows + knowledge graphs: the serialization format alone doubled my multi-hop accuracy (benchmarked 10 formats)

DGX agent

Running local models means every token counts — an 8K or 16K window fills up fast when you're stuffing graph context into prompts for RAG. I benchmarked 10 graph serialization formats (JSON, GraphML,

model-releasesr-localllama

27 Jul 2026

The Lift Spectrum: How Measurement-to-Space Adaptivity Shapes Robustness in Image-Free Single-Pixel Sensing

model-releasesarxiv-cs-cv

arXiv:2607.22077v1 Announce Type: cross Abstract: Single-pixel sensing encodes a scene as a short sequence of coded measurements, and image-free methods infer the task directly from that sequence. Rem

Time-Reversed Imaging: A Multimodal Benchmark and Framework for Reconstructing Past Human-Environment Interactions

model-releasesarxiv-cs-cv

arXiv:2607.22352v1 Announce Type: new Abstract: We introduce time-reversed imaging, a new paradigm that infers what just happened in a scene from fading multimodal traces. Instead of extrapolating or

Very cool paper from Microsoft. The idea is to train agents on replayed teacher trajectories instead of live environment rollouts. On-policy…

Very cool paper from Microsoft. The idea is to train agents on replayed teacher trajectories instead of live environment rollouts. On-policy distillation for agentic tasks is expensive because every u

safetydair-ai--x

Agents

VTM-Nav: Harnessing Cross-Episode Experience for Object-Goal Navigation with Hierarchical Visual-Topological Memory

arXiv:2607.14514v2 Announce Type: replace Abstract: Training-free ObjectNav agents increasingly use vision-language models (VLMs), yet typically discard acquired scene knowledge after each request. We

agentsarxiv-cs-cv

We are releasing PerceptionBench, a benchmark that isolates visual perception and evaluates it as a set of atomic capabilities - discovered …

model-releaseskimi-moonshot--x

We are releasing PerceptionBench, a benchmark that isolates visual perception and evaluates it as a set of atomic capabilities - discovered from how today's models fail, rather than defined in advance

We could really use Qwen3.8 in 27B, 35B, 122B and 397B sizes

Instead of 2T+ models, continuing to release highly capable small to medium size LLMs would really help to keep this community vibrant. Hardly anyone can even dream of running the recent 1.5-2T+ beast

16 bit better than lower quants for Qwen3.6-27B

I am writing a fairly complex C++ windows MFC application. I have a few 3090s and can run F16 Qwen3.6-27B with 256K context and MTP. The quality of code is exceptional with this quant vs its lower qua

90 agentic bakeoff runs: ThinkingCap vs Fable Fusion vs stock Qwen3.6-27B

Last week someone here said ThinkingCap and Fable Fusion 'really do beat the OG' for agentic work, so I ran it: 6 self-grading tasks, 5 reps, 3 models, 90 isolated runs. Tooling, since that's half the

BREAKING: A Redditor just discovered that shared Claude conversations have been showing up in public search results. The post has 4K upvotes…

model-releasesgary-marcus--x

BREAKING: A Redditor just discovered that shared Claude conversations have been showing up in public search results. The post has 4K upvotes and hundreds of comments, so this is spreading fast. Here's

Harness showdown: Claude Code vs OpenCode vs Pi with DeepSeek V4 Flash

I ran DeepSeek V4 Flash through Claude Code, OpenCode and Pi on my own benchmark, and the quality came out basically the same across all three while the time and tokens spent was wildly different. Cla

Has anyone compared pre-training, SFT/LoRA and reinforcement post-training on Qwen3.6-27B?

Qwen3.6-27B: SFT vs continued pre-training vs RL? I’m interested in adapting Qwen3.6-27B, but I’m increasingly unsure whether conventional SFT/LoRA is the best route if the goal is to add a capability

Tutorials

Simply untrue. The Economist asked Musk about a wide range of things, including AI and robotics. Watch it here for yourself: https://www.eco…

Simply untrue. The Economist asked Musk about a wide range of things, including AI and robotics. Watch it here for yourself: https://www.economist.com/insider/the-insider/an-interview-with-elon-musk?u

tutorialsgary-marcus--x

Anthropic launches Claude Opus 5 with efficiency, safety improvements

model-releasessiliconangle

Anthropic PBC today rolled out a large language model called Claude Opus 5 to its chatbot service and developer platform. The company says the LLM approaches the output quality of its top-end Mythos 5

Benchmarks: TensorSharp vs. llama.cpp

Cuda and Vulkan Benchmark: TensorSharp vs. llama.cpp I would like to share my latest open source local Unsloth (GGUF) LLM inference engine and applications. It supports many models from Unsloth, like

Agents

Diffusion LLMs can now handle real agentic work. LLaDA 2.2 is the first large-scale diffusion LLM built to operate as a real agent, planning…

Diffusion LLMs can now handle real agentic work. LLaDA 2.2 is the first large-scale diffusion LLM built to operate as a real agent, planning, calling tools, and self-correcting across long multi-turn

agentsdair-ai--x

Ollama Qwen3.6:35b randomly stops outputting tokens

RTX 4070, 32gb system ram, Linux. NVIDIA-SMI 610.43.03, KMD Version: 610.43.03, CUDA UMD Version: 13.3 Systemd service modifications: [Service] Environment='OLLAMA_HOST=0.0.0.0:11434' Environment='OLL

Very happy to support this on behalf of Google. We have long benefited from open source, are big contributors to open source and in fact hav…

model-releasesclem-delangue--x

Very happy to support this on behalf of Google. We have long benefited from open source, are big contributors to open source and in fact have consistently made open weights models with Gemma available

Applications

A Graph Neural Network approach to zero-shot Digital Twins

arXiv:2607.20535v1 Announce Type: cross Abstract: Traditional Predictive Digital Twins often remain geometrically rigid, requiring extensive retraining or fine-tuning whenever the underlying physical

applicationsarxiv-cs-ai

Achieving Text-based Person Retrieval with Any Granularity

model-releasesarxiv-cs-cv

arXiv:2607.21057v1 Announce Type: new Abstract: Text-based person retrieval faces a critical but under-explored challenge: the inherent uncertainty of query granularity in real-world scenarios. This p

Agents

Agent-Guided Relational Concept Discovery: Toward Interpretable Surgical Margin Assessment

arXiv:2607.21437v1 Announce Type: new Abstract: Deep learning models can effectively use Rapid Evaporative Ionization Mass Spectrometry (REIMS) data for surgical margin assessment. However, their clin

agentsarxiv-cs-ai

AISE-Bench: A Full-Cycle Curated Benchmark for Information Seeking on Academic Knowledge Graphs

arXiv:2607.20498v1 Announce Type: new Abstract: Large language models (LLMs) augmented with tools are emerging as autonomous agents capable of using Web engine, APIs, and code to solve complex, long-h

Am I supposed to interpret that it's better than Fable 5 from the benchmarks? Or it's ~close but cheaper?

model-releasesjerry-liu--x

Am I supposed to interpret that it's better than Fable 5 from the benchmarks? Or it's ~close but cheaper? Introducing Claude Opus 5. It's a thoughtful and proactive model that comes close to the front

Local Ai

AMD targets AI PCs to curb agentic AI costs as enterprises rethink cloud token economics

As AI moves beyond chatbots toward autonomous agents, attention is shifting enterprise AI PCs as a new layer of AI infrastructure. That transition is driving demand for hardware and software designed

local-aisiliconangle

An Evaluation Framework for Structured Audio Captions Validated by Controlled Perturbations

arXiv:2607.21424v1 Announce Type: new Abstract: Recent advancements in automated audio captioning (AAC) have shifted from monolithic sentence generation toward structured formats that explicitly disen

researcharxiv-cs-cl

Announcing Fugu-Ultra v1.1 🐡 We’ve been thrilled by the reception to the Fugu model family. Thanks to everyone who tried it, shared feedbac…

model-releasesdavid-ha--x

Announcing Fugu-Ultra v1.1 🐡 We’ve been thrilled by the reception to the Fugu model family. Thanks to everyone who tried it, shared feedback, and trusted Fugu with real work. Today, we’re releasing Fu

ArbiGraph: Arbitrarily Scalable Verifiable Task Graphs for Evaluating Context Management

arXiv:2607.20764v1 Announce Type: new Abstract: We introduce ARBIGRAPH, a benchmark generator for evaluating whether tool-assisted language agents can retain, update, compose, and discard task-relevan

AREX: Towards a Recursively Self-Improving Agent for Deep Research

arXiv:2607.21461v1 Announce Type: new Abstract: Deep research requires agents to find answers that jointly satisfy multiple constraints. Discovering such answers is costly, whereas verifying a candida

AttriMem: Attribution-Guided Process Feedback for Agent Memory Learning

arXiv:2607.21106v1 Announce Type: new Abstract: Effective memory is crucial for LLM agents, yet constructing it effectively remains challenging. A memory-construction policy decides what information t

safetyarxiv-cs-ai

Automated Synthesis and Adversarial Validation of Executable Causal Research Pipelines

model-releasesarxiv-cs-lg

arXiv:2607.21173v1 Announce Type: new Abstract: While automated research systems promise to accelerate empirical analysis, they are prone to silent failures: instances in which analysis code executes

Autonomous disproofs of the sum-product conjecture over mathbb R with GPT-5.5 Pro

arXiv:2607.20525v1 Announce Type: new Abstract: OpenAI's recent disproof of the Erdos unit distance conjecture marked a milestone for AI in mathematics. It also inspired another breakthrough: a human

Autonomous Topology Mutation: Safe Runtime Restructuring for Multi-Agent LLM Systems with Capability, State, and Shadow Invariants

arXiv:2607.20488v1 Announce Type: new Abstract: Multi-agent LLM frameworks typically fix their team topology at boot time. When an individual agent becomes overloaded at runtime, for example by mixing

Axolotl3D: a Unified Framework for Faithful 3D Shape Completion

arXiv:2607.20660v1 Announce Type: new Abstract: Recent 3D generative models produce high-quality geometry from a single image using large-scale priors and diffusion architectures. However, they assume

safetyarxiv-cs-cv

Can LLMs solve mazes?

https://reddit.com/link/1v5rvuq/video/bgmwc754i9fh1/player My goal was to create a benchmark to measure the spatial awareness and memory of models. Eventually, I came up with the simple idea of a maze

Can Valence Reflect Morality in Natural Language? A Preliminary Annotation Study

arXiv:2607.20461v1 Announce Type: cross Abstract: Present implementations of artificial intelligence (AI) ethics do not adequately take feelings, or affect, into account. If AI should be aligned with

researcharxiv-cs-ai

Capital Markets LLM Reliability Score (CM-LRS): From Plausible to Bankable

model-releasesarxiv-cs-cl

arXiv:2607.21340v1 Announce Type: new Abstract: In capital-markets workflows the question is rarely whether a large language model can produce a fluent draft, but whether the draft is bankable: defens

Cautious optimism for deep parameterized quantum circuits

arXiv:2607.21409v1 Announce Type: cross Abstract: A central challenge in quantum machine learning is understanding the scaling behavior of parameterized quantum circuits (PQCs). In particular, it rema

researcharxiv-cs-lg

Local Ai

Concept Concentration for Faithful Representation Intervention

arXiv:2505.18672v2 Announce Type: replace Abstract: Representation intervention aims to localize and modify the representations that encode the underlying concepts in large language models (LLMs) to e

local-aiarxiv-cs-lg

CultureTalk-ID: A Multi-Task Dialogue Benchmark for Cultural Commonsense in Indonesian Local Languages

model-releasesarxiv-cs-cl

arXiv:2607.21016v1 Announce Type: new Abstract: Culture is lived through conversation, yet existing Indonesian cultural commonsense benchmarks evaluate LLMs on short and isolated prompts, stripping aw

DFAH-Bench: Benchmarking Observable Agent Instability in Financial Decision-Making

arXiv:2607.20491v1 Announce Type: new Abstract: Standard evaluation benchmarks measure what a tool-using agent decides, not whether it arrives at that decision through the same process each time. We i

Explanation-Based Runtime Verification for Trustworthy ML-driven Optical Networks

arXiv:2607.20675v1 Announce Type: new Abstract: Machine learning (ML) models are increasingly integrated into optical network automation frameworks to support tasks such as failure management, perform

researcharxiv-cs-lg

Extened garlic to run Qwen3.5 35B A3B float8 at 55 tok/s on RTX 5060 Ti

In a previous post (https://www.reddit.com/r/LocalLLaMA/comments/1utefpr/running_qwen3_30b_a3b_at_50_toks_on_rtx_5060_ti/) there seemed to be great demand for bringing in Qwen3.5 35B. Some Gated Delta

Fizgig Krea 2 training features update

https://github.com/shootthesound/Fizgig Intelligent trainer - Per-image loss tracking with self-adapting training runs — every image gets its own verdict (easy / suspect / stuck / exhausted) and its o

safetyr-stablediffusion

For my first post, I’m sharing a letter @NVIDIA signed on why open models matter. AI will transform every industry, power every company, and…

For my first post, I’m sharing a letter @NVIDIA signed on why open models matter. AI will transform every industry, power every company, and be built by every country. Open models strengthen safety an

safetyclem-delangue--x

GeoThreat: Transferable Targeted Adversarial Attacks on Large Vision-Language Models for Remote Sensing Image Interpretation

arXiv:2607.21036v1 Announce Type: new Abstract: Adversarial attacks against large vision-language models (LVLMs) serve as an effective means of assessing their robustness in cross-modal semantic under

researcharxiv-cs-cv

Hybrid MKNF with Classical Negation in the Rule Component

arXiv:2607.21202v1 Announce Type: cross Abstract: Hybrid MKNF knowledge bases under the well-founded semantics integrate Description Logics with Logic Programming. However, they do not support classic

safetyarxiv-cs-ai

HyWorldVLA: A Vision-Language-Action Model with Hybrid World Modeling for Autonomous Driving

arXiv:2607.20988v1 Announce Type: cross Abstract: Vision-Language-Action (VLA) models augmented with world modeling represent a promising paradigm for end-to-end autonomous driving. While pixel-level